Professional Documents
Culture Documents
ABSTRACT
Predictions of the mean and covariance matrix of summary statistics are critical for confronting cosmological theories with
observations, not least for likelihood approximations and parameter inference. The price to pay for accurate estimates is
the extreme cost of running 𝑁-body and hydrodynamics simulations. Approximate solvers, or surrogates, greatly reduce the
computational cost but can introduce significant biases, for example in the non-linear regime of cosmic structure growth. We
propose "CARPool Bayes", an approach to solve the inference problem for both the means and covariances using a combination
of simulations and surrogates. Our framework allows incorporating prior information for the mean and covariance. We derive
closed-form solutions for Maximum A Posteriori covariance estimates that are efficient Bayesian shrinkage estimators, guarantee
positive semi-definiteness, and can optionally leverage analytical covariance approximations. We discuss choices of the prior
and propose a simple procedure for obtaining optimal prior hyperparameter values with a small set of test simulations. We test
our method by estimating the covariances of clustering statistics of GADGET-III 𝑁-body simulations at redshift 𝑧 = 0.5 using
surrogates from a 100-1000× faster particle-mesh code. Taking the sample covariance from 15,000 simulations as the truth, and
using an empirical Bayes prior with diagonal blocks, our estimator produces nearly identical Fisher matrix contours for ΛCDM
parameters using only 15 simulations of the non-linear dark matter power spectrum. In this case the number of simulations is so
small that the sample covariance would be degenerate. We show cases where even with a naïve prior our method still improves
the estimate. Our framework is applicable to a wide range of cosmological and astrophysical problems where fast surrogates are
available.
Key words: large-scale structure, cosmological simulations, 𝑁-body, covariance
1 INTRODUCTION in terms of redshift, sky area, volume, etc,... We then use the sam-
ples to compute the unbiased and positive definite sample covariance
To study the large-scale structure of the universe and cosmic growth
estimator, but getting a reliable estimate requires many realizations,
history in the era of data-driven cosmology, one needs to accurately
especially if we need the precision matrix for the estimation of pa-
model the statistical properties of observables in order to infer cosmo-
rameter confidence bounds.
logical parameters constraints from surveys. The covariance matrix
𝚺 of a summary statistics vector, such as the matter power spectrum To reduce the computational cost of generating simulation samples
across different wave-numbers, and most importantly its inverse — various parallel, distributed-memory 𝑁-body solvers have been de-
the precision matrix—are paramount to extracting low-dimensional veloped sometimes with GPU-acceleration ( Springel (2005) (GAD-
summaries, building inference frameworks or computing likelihood GET), Ishiyama et al. (2009) (GreeM), Warren (2013) (2HOT),
approximations from mock catalogues (Heavens et al. (2000); Eifler, Harnois-Déraps et al. (2013) ( CUBEP3 M), Garrison (2019) (Aba-
T. et al. (2009); Takahashi et al. (2009); Harnois-Déraps et al. (2012); cus), Habib et al. (2016) (HACC), Potter et al. (2017) (PKDGRAV3),
Dodelson & Schneider (2013); Harnois-Déraps & Pen (2013); Taylor Yu et al. (2018) and Cheng et al. (2020) (CUBE)). Relying solely
& Joachimi (2014); Percival et al. (2014); Blot et al. (2014); Joachimi on massively parallel computing to tackle next-generation obser-
& Taylor (2014); Alsing & Wandelt (2018); Harnois-Déraps et al. vational datasets appears impractical given our time, memory, and
(2019); Hikage et al. (2020); Wadekar et al. (2020); Giocoli et al. energy resources since thousands of simulations are needed to pro-
(2021)). duce sufficiently accurate cosmological parameter constraints (see
The most trusted yet costly method to compute the covariance for instance Blot et al. (2016)).
matrix of large-scale structure clustering statistics consists in gener- For this reason, cosmologists have been searching for alternatives
ating mock realizations of survey observables with intensive 𝑁-body to running a large number of 𝑁-body simulations for a particular
simulations – or even hydrodynamical simulations for certain appli- cosmological model.
cations – that mimic the conditions of observational data sampling On the theoretical side, analytical computations give covariance
Figure 1. Illustrating the power of Bayesian control variates using the confidence contours of the cosmological parameters computed using the Fisher matrix
based on the estimated matter power spectrum covariance matrix. The "truth" designates the confidence regions (black) from the sample covariance matrix of
15,000 𝑁 -body simulations, and the parameter means are set to the ΛCDM model used in the simulations. The contours are overlapped nearly perfectly by the
light blue when the covariance in the Fisher matrix is computed using only 15 simulations with our CARPool Bayes MAP estimator (10 simulations and 5 for
setting a prior hyperparameter, see section 3.5). The sample covariance (ML) estimator based on many more simulations than ours gives less accurate contours.
Contours based on 3100 COLA surrogates alone are rotated and too small showing that the surrogates alone are inaccurate. Detailed discussion in the text and
in section 4.
compute eigenvalues will dominate the precision matrix and impact parameter
parameter constraints (Taylor et al. (2013), Blot et al. (2016)).
𝑛𝑠
𝛾 ∑︁ Now we add surrogates. The goal is to build a Bayesian model
𝚺𝒔𝒔 =
b (𝒔𝑖 − 𝒔¯) (𝒔𝑖 − 𝒔¯)𝑻 (3) for the covariance of the simulations but including whatever in-
𝑛𝑠
𝑖=1 formation is provided by the surrogates. The set of surrogates 𝒓 𝑗 ,
𝑛𝑠
1 ∑︁ 𝑗 = 1, . . . , 𝑛𝑠 + 𝑛𝑟 comprises 𝑛𝑠 samples that are paired with the
𝒔¯ = 𝒔𝑖 , simulations, i.e., they were computed using the same random num-
𝑛𝑠
𝑖=1 bers, and 𝑛𝑟 additional unpaired surrogates. We combine pairs of
simulations and surrogates into a single vector
the Maximum-Likelihood (ML) estimator given a Multivariate Nor-
mal (MVN) likelihood function when 𝛾 = 1. To get an unbiased
estimator, we use Bessel’s correction factor 𝛾 = 𝑛𝑠 /(𝑛𝑠 − 1) in the
ML estimator for the covariance. Equation (3) needs many samples
to provide a high-quality estimate: as a matter of fact, the conver- 𝑻
gence of the smallest eigenvalues is slow (Bai & Yin 1993) and these 𝒙 ≡ 𝒔, 𝒓 (4)
P (𝒔∗ |𝒓 ∗ , 𝚺) = 𝑀𝑉 𝑁 ( 𝝁 𝒔∗ |𝒓 ∗ , 𝚺 𝒔∗ |𝒓 ∗ ) (6) A convenient prior to choose for the block covariance 𝚺, with 𝑃 ≡
−1 ∗ 𝑝 𝑠 + 𝑝 𝑟 , is the Inverse-Wishart (W −1 ) prior with hyperparameters
𝒔b∗ = 𝝁 𝒔∗ |𝒓 ∗ = 𝚺 𝑠𝑟 𝚺𝑟𝑟 (𝒓 − 𝝁𝒓 ) + 𝝁 𝒔 + , the scale matrix, and 𝜈, the number of degrees of freedom.
𝚿 ∈ S𝑃
𝚺 𝒔∗ |𝒓 ∗ = (𝚺/𝚺𝑟𝑟 ) With 𝑛 𝑝 ≡ 𝜈 + 𝑃 + 1 then
The regression matrix of 𝒔 given 𝒓 will appear from now on as det(𝚿) 𝜈/2 1 −1
W −1 (𝚺|𝚿, 𝜈) = 𝜈
det(𝚺) −𝑛 𝑝 /2 𝑒 − 2 tr(𝚿𝚺 ) (11)
−1 𝑃( 2)
2𝜈 𝑃/2 Γ
𝑩 ≡ 𝚺 𝑠𝑟 𝚺𝑟𝑟
𝚿𝒔𝒔 𝚿𝒔𝒓
For legibility, and without loss of generality, we will write all 𝚿≡ ,
𝚿𝒓 𝒔 𝚿𝒓𝒓
random vectors as zero-mean in the derivations such that for any
sample 𝑖 where Γ 𝑃 is the multivariate Gamma function. W −1 (𝚺|𝚿, 𝜈) has
𝒙𝑖 ← 𝒙𝑖 − 𝝁𝒙 . mode 𝚿/𝑛 𝑝 for 𝑛 𝑝 > 2𝑃. Its mean 𝚿/(𝑛 𝑝 − (2𝑃 + 2)) exists if
𝑛 𝑝 > 2𝑃 + 2. In our problem, for any prior P (𝚺), the mode of the
The final equations serving as numerical recipes will include the
means explicitly.
With these notations, we now turn to inferring the simulation block 3 Anderson (1957) derived the same ML estimator by integrating out the 𝒔 ∗
of the covariance 𝚺𝒔𝒔 with the help of surrogates, given (multiple in Eq. (7) to obtain the marginal likelihood for the observed samples only
realisations of) 𝒙 and 𝒙 ∗ . 𝑛𝑠
∑︁
− 2 ln L ( {𝒙 }, {𝒓 ∗ } |𝚺) = 𝑛𝑠 ln [det (𝚺) ] + 𝒙𝑻 −1
𝑖 𝚺 𝒙𝑖
𝑖=1
𝑛𝑟
3.2 Maximum-likelihood solution with surrogates ∑︁
+ 𝑛𝑟 ln [det (𝚺𝒓𝒓 ) ] + 𝒓 ∗𝑗 𝑻 𝚺𝒓𝒓 −1 𝒓 ∗𝑗 + 𝑐𝑚 ,
In a Gaussian model, the log-likelihood of 𝑛𝑠 independent and 𝑗=1
identically distributed (iid) samples of 𝒙 and 𝑛𝑟 iid samples 𝒙 ∗ of with 𝑐𝑚 the remaining constant of the model with missing 𝒔 ∗ .
3.3.2 MAP with prior on the regression parameters A common form adopted as a target for shrinkage estimates of co-
variance matrices is the "identity" prior: the auto-covariance of sim-
A different approach is to solve the MAP for the parameters that allow ulations and surrogates are proportional to identity matrices and the
to estimate 𝚺𝒔𝒔 = 𝚺𝒔 |𝒓 + 𝑩𝚺𝒓𝒓 𝑩𝑻 , that is to say we use a prior for
the joint distribution P (𝑩, 𝚺𝒔 |𝒓 , 𝚺𝒓𝒓 ) which is a reparametrization
4
of the 𝑝 𝑠 ( 𝑝 𝑠 + 1)/2 + 𝑝 𝑟 ( 𝑝 𝑟 + 1)/2 + 𝑝 𝑠 𝑝 𝑟 parameters of P (𝚺). For We know that for two random vectors 𝒙 and 𝒚 with 𝒚 = ℎ (𝒙), if ℎ is differ-
that, we need the properties of the blocks of a covariance sampled entiable, then for probability distributions P𝑦 (𝒚) = P𝑥 (𝒙) × det 𝑱 ℎ−1 (𝒚)
from a W −1 (𝚺|𝚿, 𝜈) distribution. A quick outline of the derivation where 𝑱 is the Jacobian matrix. So under a reparametrization, the two distri-
appears in Appendix B. With 𝑩𝚿 ≡ 𝚿𝒔𝒓 𝚿𝒓𝒓 −1 we get butions have no reason to peak at the same coordinates.
𝚿emp ≡ 0 0
®
𝜎𝑠2𝑝𝑠
MAP
Compute b 𝚺𝒔𝒔 (𝑛 𝑘𝑝 ) using equations (14) to (17).
𝜌 𝑝𝑠 𝜎𝑟 𝑝𝑠 𝜎𝑠 𝑝𝑠 ®® 2
0 Compute the MVN likelihood L {𝒔}𝑡𝑒𝑠𝑡 |𝚺𝒔𝒔 (𝑛 𝑝 ) .
® 3
𝜎𝑟21
®
𝑻 end
®
4
𝚿𝒔𝒓
®
..
Determine 𝑛★𝑝 = argmax N𝑝 L {𝒔}𝑡𝑒𝑠𝑡 |𝚺𝒔𝒔 (𝑛 𝑝 )
. ®
® 5
0
®
𝜎𝑟2𝑝𝑟 6 return b 𝚺MAP (𝑛★ ); E
𝒔𝒔 𝝁MAP (𝑛★ )
« ¬ 𝑝 𝒔 |𝒓 𝑝
(24)
The computation of each 𝜎𝑠2𝑖 , 𝜎𝑟2𝑖 and 𝜌𝑖 𝜎𝑠𝑖 𝜎𝑟𝑖 is the same as from
the "identity" prior above.
While having a very simple structure, we can see this prior as a 3.6 Correction factor for the precision
more adapted correction of the eigenvalues of the block matrix 𝚺
To compute confidence bounds of the cosmological parameters in
based on the data, whereas 𝚿id adds the same amount of correction
the context of a likelihood-analysis, we need to invert the covari-
on all the eigenvalues, regardless of the statistics at hand.
ance matrix estimate. We briefly explain the correction used for the
standard sample covariance.
3.5 (Cross-)validation to choose the prior hyperparameter 𝑛 𝑝
The hyperparameter 𝜈 (through 𝑛 𝑝 = 𝜈 + 𝑝 𝑠 + 𝑝 𝑟 + 1) in equation 3.6.1 Classical result for the sample covariance
(11) will be seen to determine the weight attributed to the prior in the
MAP
It is well-known that taking the inverse of the bias-corrected version
closed-form solutions for b 𝚺𝒔𝒔 . For different statistics, and in terms
of the Maximum-Likelihood estimator from equation (10), i.e 𝛾b 𝚺𝒔𝒔
of the maximum number of simulations 𝑛𝑠 one is able to run, varying
where 𝛾 ≡ 𝑛𝑠 /(𝑛𝑠 − 1), results in a biased estimator of the precision
𝑛 𝑝 via 𝜈 can significantly impact the quality of the covariance, as we
will discuss in section 4.
We propose retaining a small set {𝒔 𝑡𝑒𝑠𝑡 } of test simulations such that 5 We compared this to using 𝐾 -fold cross-validation but found no signif-
𝑛𝑠 = 𝑛cov test cov
𝑠 + 𝑛 𝑠 , where 𝑛 𝑠 plays the role of the 𝑛 𝑠 of the paired set icant impact on the determination of the optimal 𝑛 𝑝 comparatively to just
in equations (13) to (22). evaluating the likelihood once without splitting the data.
b𝒔𝒔 = 𝑛𝑠 − 𝑝 𝑠 − 2 b
𝑷 𝚺𝒔𝒔
−1
(25) 4.1.3 Post-processing of snapshots
𝑛𝑠 − 1
We chose, for this study, to include what in the cosmology lit- To extract the summary statistics from our L-PICOLA snapshots, we
erature is referred to as the "Hartlap factor" to the inverse of the used the exact same code modules and parameters used to compute
bias-corrected sample covariance of simulations summary statistics the clustering statistics available in the Quijote data outputs. There-
(including the truth using 15,000 simulations). fore, the simulation and surrogate summary statistics have the same
dimension 𝑝 𝑠 = 𝑝 𝑟 . We transform the snapshots in density contrast
fields with the Cloud-In-Cell (CiC) algorithm. For the matter power
3.6.2 For the MAP estimates spectra and the correlation functions, we used the Python3 module
Pylians3 7 , For the bispectra, the results of which appearing in Ap-
Our MAP estimates derived in section 3.3.1 is constructed to ensure pendix C3, the post-processing code is pySpectrum8 . More details
that the result will be a symmetric positive semi-definite matrix. As can be found in CWAV20.
a consequence, we lose formal unbiasedness but gain dramatically
improved estimates according to multiple criteria, as discussed in
section 4. If unbiasedness of the covariance estimate is important the 4.2 The CARPool Bayes estimator and results on clustering
method in CW21can be used. statistics
The following tests of the Bayesian covariance estimation approach in
this paper use the sample covariance matrix with all the simulations
4 NUMERICAL EXPERIMENTS ON ΛCDM SIMULATION we have (𝑛𝑡𝑟 𝑢𝑡 ℎ = 15, 000) as the "truth" to compare with other
𝑠
STATISTICS estimates. Within the main part of this paper we only present a subset
4.1 Simulation and surrogate data of the estimators that gave the best match in terms of parameters
constraints with respect to the truth.
The simulation and surrogate solvers we use are identical to those In particular, we use the MAP estimator from section 3.3.1 with
CW21 and CWAV20. We recall the main points here for conve- the "diagonal" empirical Bayes prior 𝚿emp , equation (24), estimated
nience. For more details please refer to CWAV20. The solvers evolve on the paired set of 𝑛𝑐𝑜𝑣 simulations and surrogates. All our MAP
𝑠
Np = 5123 Cold Dark Matter (CDM) particles in a box volume of covariance estimates with simulations and surrogates use the optimal
3
1000 ℎ−1 Mpc . The simulation-surrogate sample pairs take the 𝑛 𝑝 determined through the process described in Algorithm 1 with
a small number of test simulations. We display the total number
same Second-order Lagrangian perturbation theory (2LPT) initial
of simulations used for each covariance matrix estimate as 𝑛𝑠 =
conditions at starting redshift 𝑧 𝑖 = 127.0.
𝑛cov test
𝑠 + 𝑛𝑠 .
In the following, we will refer to this approach as the "CARPool
4.1.1 𝑁-body solver Bayes" estimator.
We find that the alternative estimator written in terms of the regres-
We downloaded the 𝑁-body snapshots clustering statistics from the sion parameters, section 3.3.2, performs comparatively poorer than
publicly available Quijote simulation suite6 (Villaescusa-Navarro the CARPool Bayes estimator. We show an example on the power
et al. 2020). The solver for all the simulations is the TreePM code spectrum covariance and discuss the reasons for this in Appendix
GADGET-III built upon the previous version GADGET-II by Springel C2. Briefly summarized, this estimator requires using a proper prior
(2005). The force mesh grid size to solve the comoving Poisson equa- and therefore affords us less flexibility in choosing the weight of the
tion at each timestep is Nm = 1024. In the following, we will use prior. It therefore tends to give covariance estimates that are more
the sample covariance of all 15,000 available realizations of the fidu- sensitive to the choice of the prior parameters.
cial cosmology as the simulation "truth", or more precisely the best The plan for the remainder of this sections is as follows: we will
covariance estimate we have access to. first present the power spectrum results in more details that were
already partially described in section 2.
4.1.2 Surrogate solver Then, we turn to the real space 2-point correlation function. This
is an interesting case because it illustrates the power of limiting the
We generate the fast surrogate samples with The COmoving La- range of the estimator to the set of all positive definite covariance
grangian Acceleration (COLA) method from Tassev et al. (2013) (see matrices, a feature of the Bayesian version of CARPool. The unbiased
also Leclercq et al. (2020)), which allows generating approximate CARPool approach to the covariance matrix in CW21 failed to yield
gravitational 𝑁-body outputs using a smaller number of timesteps positive-definite covariance estimates for this application in spite of a
than our simulation code. The principle of COLA is to add residual significant reduction of variance for the covariance matrix individual
displacements, computed with a Particle-Mesh (PM) 𝑁-body solver, elements.
to the trajectory given by analytical LPT approximations (usually For a complete comparison with CW21, we also computed results
first- or second-order). Izard et al. (2016) proposed tests of the accu- on the bispectrum covariance matrix. Since these show similar, large
racy and computational cost of COLA against 𝑁-body simulations at improvement over the CW21 approach as for the power spectrum,
different redshifts and with different timestepping parameters. Like we relegate details to Appendix C3.
in CWAV20 and CW21, we used the parallel MPI implementation
7 https://github.com/franciscovillaescusa/Pylians3
6 https://quijote-simulations.readthedocs.io/en/latest/ 8 Available at https://github.com/changhoonhahn/pySpectrum
𝑝 𝑠 = 158 linearly space bins. Note that the power spectrum is not
compressed unlike in CWAV20 and CW21, making the covariance
estimation tasks more difficult.
Figure 1 shows that using only 𝑛𝑠 = 10 + 5 simulations with paired
surrogates and an additional set of surrogate samples, we get confi-
dence bounds for the cosmological parameters which are very close
to the ones given by the "true" sample covariance using 15,000 sim-
ulations (for 𝑝 𝑠 = 158). This result is all the more encouraging that
with only 10 simulations we would get a sample covariance of rank
at most 9. In other words, we can see the small set of simulations in Figure 2. Illustration of step 5 of Algorithm 1 for the matter power spectrum
the "CARPool Bayes " estimate a correction to the eigenvalues and example: for fixed 𝑛𝑠 = 𝑛𝑠cov + 𝑛𝑠test and fixed prior 𝚿emp , we compute the
eigenvectors of the precision matrix computed from a biased but cor- MAP
𝚺𝒔𝒔
likelihood of b (𝑛 𝑝 ) on test simulation samples.
related surrogate. We also show in Appendix C1 the relatively small
gain, in terms of closeness of the parameters confidence contours to
the truth, of running 𝑛𝑠 = 40 + 10 simulations for comparison. other estimates and avoids the characteristic underestimation of small
Here we examine the procedure to determine the best 𝑛 𝑝 for a eigenvalues for covariance matrices estimated from a small number
given 𝑛𝑠 and 𝚿 in Figure 2. There are several points to notice here: of samples.
(i) For 𝑛 𝑝 ≈ 1, especially for 𝑛𝑠 ≥ 𝑝 𝑠 + 1 (when the sample
covariance can be full-rank), the likelihood on test data rapidly in-
4.2.2 Matter correlation function covariance
creases. It shows for this case that a minimum of "regularization"
brought by the prior greatly improves the estimate of 𝚺𝒔𝒔 . The example of the two-point matter correlation function 𝜉 (𝒓) for 𝒓 ∈
(ii) Around the empirical 𝑛★𝑝 , the likelihood is rather flat and [5.0, 160.0] ℎ−1 Mpc (𝑝 𝑠 = 159) is of particular interest in our study.
slowly decreases when 𝑛 𝑝 > 𝑛★𝑝 . In other words, once a certain With the variance reduction approach in CW21 for the covariance
threshold of "improvement" is reached with 𝑛 𝑝 , misestimating 𝑛 𝑝 matrix, we found no improvement over the standard estimator. While
does not radically worsen the estimate of 𝚺𝒔𝒔 . unbiased and strongly reducing the errors of all individual elements
of the covariance matrix the resulting matrix failed to be positive-
In Figure 3, we visualize the estimated covariance matrices (top
definite. This means that no estimate for the precision matrix could
row) and their inverse (bottom row). For the "CARPool Bayes" esti-
be obtained, as would be required to derive Fisher matrices or for a
mate with the prior 𝚿emp , i.e. our "headline" estimate with 𝑛𝑠 = 10+5
likelihood approximation to derive parameter constraints.
that gives the confidence bounds in Figure 1, we notice some structure
As we can observe in Figure 5, the structure of the covariance
in the covariance due to the small number of simulations. The close-
is particular, with a band of high-magnitude covariances around the
ness to the truth of the "CARPool Bayes" covariance with very few
diagonal of variances. As a result, the precision estimate based on
simulations is particularly visible for the structure of the precision
the standard sample covariance estimator is very noisy for 𝑛𝑠 =
matrix. It can be seen that at low 𝑘, where the correlation between
200, which we compare with our estimate including surrogates from
surrogates and simulations is particularly high, the CARPool Bayes
Algorithm 1, with 𝑛𝑠 = 160 + 20. Looking at the precision matrices
estimate (and the Maximum Likelihood estimate without the prior)
(bottom row) would indicate a significantly better recovery of the
is significantly less noisy than the standard estimator even though it
structure of the true precision.
uses an order of magnitude less simulations.
In terms of cosmological parameter forecast constraints, as shown
In Figure 4, we compare the covariance estimates to the large-
in Figure 6 , we get a slight improvement with respect to the sample
sample "truth" in the spectral domain, showing the eigenvalues and
covariance matrix (and the precision including the Hartlap factor),
the co-diagonalization coefficients 9 .
but not nearly as large as for the matter power spectrum. The CARPool
At the top, we show the ordered eigenvalues ratio of each matrix. A
Bayes estimate with 𝑛𝑠 = 160 + 20 produces bounds for Ω𝑚 , 𝑛𝑠 and
ratio far from 1, and especially close to zero for the smallest eigenval-
𝜎8 that are closer to the truth than with the sample covariance with
ues as for the standard sample covariance, indicates a very poor con-
𝑛𝑠 = 180. But the confidence regions for Ω𝑏 and ℎ are not improved.
ditioning of the matrix. At the bottom, we see the co-diagonalization
Similarly to the previous section, in Figure 7, the "CARPool Bayes"
coefficients. A horizontal line at 1 would indicate that the matrices
estimator raises up the smallest eigenvalues – as well as the smallest
are identical. The CARPool Bayes estimate clearly outperforms the
"co-diagonalization" coefficient – contrarily to the ML solutions with
and without surrogates where they are close to 0.
9 For 𝑨 and 𝑩 two 𝑝 × 𝑝 real symmetric matrices, if 𝑨 is positive definite, The wide band of correlations visible in Figure 5 indicate that
then there exists a matrix 𝑴 such that 𝑴 𝑻 𝑨𝑴 = 𝑰 𝑝 and 𝑴 𝑻 𝑩𝑴 = 𝑫 with our choice of "diagonal" prior is not optimal for this case. Choosing
𝑫 = diag(𝑑1 , . . . , 𝑑 𝑝 ). We call the 𝑑𝑖 "co-diagonalization coefficients." This a prior with a more gradual falloff of correlation from the diagonal
is a simplified statement from theorem A9.9 in Muirhead (1982). If D = 1 𝑝 would likely produce better results. Figure 8 indicates that for various
then A = B. number of simulations 𝑛𝑠 , the CARPool Bayes estimator for the
matter correlation function covariance consistently prefers low 𝑛 𝑝 vious approaches such as the sample covariance or the first-order
(i.e. prior weight) values with the "diagonal" prior from section CARPool approach in CW21 according to multiple criteria. These
3.4.2. improvements are particularly noticeable for the inverse covariance
In summary, the application to the matter correlation function, or precision matrix required for many applications such as comput-
demonstrates that the CARPool Bayes estimator is guaranteed to ing Fisher matrices, or for the Gaussian likelihood approximations
produce positive definite matrices. It visibly improves the structure frequently used for parameter estimation.
of the precision matrix (Figure 5) and the relative errors of the Our Bayesian approach combines estimations for the both the mean
small eigenvalues (Figure 7). This translates into some, but not all, (through the well known regression 𝝁 𝒔 |𝒓 , equation 16) and the covari-
parameter confidence bounds being closer to the truth than for the ance of simulation summary statistics using surrogates. In this paper
sample covariance based on 180 simulations. we focused on showing the results for the simulation covariance esti-
mates 𝚺𝒔𝒔 since this is the first time that the control variate approach
has been cast in a Bayesian framework for covariance estimation.
Our Bayesian approach used a multivariate Gaussian model for the
5 DISCUSSION AND CONCLUSION
simulations and surrogates and includes a conjugate Inverse-Wishart
We consider the problem of estimating the covariance matrices of distribution prior for the covariance matrix. In the generic case we
cosmological summary statistics within a Bayesian framework, when found a "diagonal" prior on the block covariance of simulation and
paired simulations and surrogates are available. surrogate summary statistics, whose diagonal elements were eval-
This study constitutes an extension of the CARPool principle, uated on simulation-surrogate pairs, section 3.4.2, to give excellent
presented in CWAV20 and applied to covariance matrices in CW21. results, especially for the matter power spectrum and bispectrum. We
Our method improves on the latter work by solving a Maximum A obtain the same confidence bounds as with the true covariance of the
Posterior optimization directly in the space of symmetric positive power spectrum with 𝑝 𝑠 = 158 bins up to 𝑘 ≈ 1.0 ℎMpc−1 with only
semi-definite matrices and allows introducing priors in analogy to 𝑛𝑠 = 10 + 5 simulations. In this case, we can think of the actual 10
frequentist shrinkage estimators. We prove that our approach, dubbed simulations of the covariance estimate as correcting the eigenspec-
CARPool Bayes, guarantees positive definite estimates, for the price trum of the well-converged covariance of the correlated surrogate
of abandoning the guarantee of unbiasedness of individual covariance that incorporates many samples.
matrix elements provided by the first order estimator described in The same outstanding gain appears for the bispectrum, as we show
CW21. in Appendix C3 for two triangle configurations. This demonstrates
By casting CWAV20 in a Bayesian framework we provide a new the superiority of the CARPool Bayes approach over CW21.
solution to covariance estimation with simulations and surrogates. Regarding the 2-point matter correlation function in real space,
We demonstrate that this estimator can strongly improve over pre- we do get positive-definite estimates by construction—this is not
Figure 4. Comparison of the CARPool Bayes covariance estimate (section 5.2 Potential for future applications in cosmology and beyond
3.3.1), the standard ML estimator, and the ML estimator combining simula- Our numerical experiments demonstrate the capability of running
tions and surrogates (section 3.2) with the large-sample "truth" in the spectral
fewer intensive simulations in order to get theoretical predictions
domain. We show ordered eigenvalue ratio at the top and co-diagonalisation
coefficients at the bottom. The CARPool Bayes estimator avoids the character- of the means and covariances of observables for next-generation
istic underestimation of small eigenvalues for covariance matrices estimated surveys. Many additional applications of these techniques remain to
from a small number of samples. See discussion in the text. be explored. The free choice of what to use as surrogates makes our
methods very broadly applicable.
Some surrogates might be useful because they are nearly free com-
putationally. A case in point would be Eulerian linear perturbation
theory for the power spectrum applied to the initial conditions of
an 𝑁-body simulation. In this case each simulation comes with the
guaranteed in CW21—and we obtain a slight improvement on the paired surrogate for free (since the initial conditions are necessary to
parameter constraints with respect to the sample covariance of sim- run the simulation in the first place) and its expectation and covari-
ulations when 𝑛𝑠 & 𝑝 𝑠 + 1 is close to the dimension of the summary ance can be computed analytically nearly for free. It could be argued
statistics. But in a case where running a high enough number (we that such automatic surrogates ought to be exploited systematically
tested 𝑛𝑠 = 300) of simulations is possible, the gain over the sample when predicting commonly used clustering statistics from simula-
covariance diminishes as 𝑛𝑠 increases, at least regarding the impact tions. A very similar application of this idea to a non-perturbative
of the matter correlation function covariance on the parameter con- statistic would be to the computation of halo number functions: apply
straints. the Press-Schechter approach to the initial conditions as a surrogate
Throughout this study, we applied the "diagonal" empirical Bayes for the mass function for a given realisation. In this example, the
prior through taking the diagonal of each block of the summary statis- classical Press-Schechter formula provides the expectation of the
tics as a concatenation of the simulation and surrogate output. Using surrogate and would reduce the variance in the number function for
the former was sufficient to demonstrate the capability of the method the largest (and rarest) clusters in the simulations, thus increasing the
for the case where we consider the problem of estimating the whole effective volume of the simulations.
block covariance 𝚺 to then extract the simulation block 𝚺𝒔𝒔 . We In other cases, the surrogates may consist of costly simulations
derived new MAP estimators in section 3.3.2 where we directly esti- that have already been run at a different set of parameters. In this
mate the regression parameters allowing to compute the simulation case it may be possible to "update" the means and covariances from
block of the covariance using 𝚺 𝑠𝑠 ≡ (𝚺/𝚺𝑟𝑟 ) + 𝚺 𝑠𝑟 𝚺𝑟𝑟 −1 𝚺 the previous simulation set to a new set of parameters by pairing a
𝑟 𝑠 from
the hypothesis of data sampled from a MVN distribution. This new small number of the existing old simulations (now surrogates) with
estimator did not provide improvement in our tests over the sample the same number of new simulations.
covariance, which we attribute to the strong prior dependence inher- The availability of perturbative results and analytical estimates,
ent in it. We leave for future studies the question whether the this the increasing need for accurate simulations to analyze current and
different parametrization can turn beneficial for cosmological survey upcoming data sets in all subfields of cosmology, and the vast param-
forecasts when theoretically motivated covariances for the prior are eter space to explore with cosmological simulations make it likely
available. that the concepts described here will continue to find powerful appli-
Figure 5. Matter correlation function covariance estimates (top) and their inverse (bottom), shown similarly to Figure 3. The "CARPool Bayes" estimates uses
𝑛𝑠 = 160 + 20 GADGET simulations.
cations. We look forward to seeing the cosmological advances that APPENDIX A: DERIVATION OF ESTIMATORS AND
CARPool will enable. PROOF OF POSITIVE DEFINITENESS USING
EXPECTATION MAXIMIZATION
A1 Expectation Maximization
In this section, we aim at showing the equivalence of the results given
by the Expectation-Maximum algorithm – which naturally comes to
ACKNOWLEDGEMENTS mind in the presence of missing samples – and the simple result
from the Maximum-Likelihood and Maximum a Posteriori problems
We warmly thank Ethan Anderes and Francisco Villaescusa-Navarro formulated in sections 3.2 and 3.3.1.
for stimulating discussion and feedback. N.C. acknowledges funding
from LabEx ENS-ICFP (PSL). B.D.W. acknowledges support by the
ANR BIG4 project, grant ANR-16-CE23-0002 of the French Agence A1.1 Iterative algorithm
Nationale de la Recherche; and the Labex ILP (reference ANR-10- The Expectation-Maximization (EM) algorithm (Dempster et al.
LABX-63) part of the Idex SUPER, and received financial state aid 1977) is an iterative technique to maximize the likelihood (or pos-
managed by the Agence Nationale de la Recherche, as part of the terior) in the presence of missing data. Briefly, it works by casting
programme Investissements d’avenir under the reference ANR-11- the problem as a sequence of simpler optimization problems. Each
IDEX-0004-02. The Flatiron Institute is supported by the Simons iteration consists of two steps: the E-step which removes the missing
Foundation. This work has made use of the Infinity Cluster hosted data from the log-likelihood by taking its expectation with respect to
by Institut d’Astrophysique de Paris. the missing data assuming the current iterates are the true values of
the parameters; and the M-step which updates the parameters by find-
ing their values that maximize the expected log-likelihood from the
E-step. We focus in this appendix on the covariance estimation; in-
cluding the solution for the estimators of the mean is straightforward
and we give the result in the main text.
DATA AVAILABILITY We recall Eq. (7) here for convenience as a starting point
− 2 ln L ({𝒙}, {𝒙 ∗ }|𝚺) = (𝑛𝑠 + 𝑛𝑟 ) ln [det (2𝜋𝚺)]
The data samples underlying this article are available through
globus.org, and instructions to reproduce the summary statis- ∑︁𝑛𝑠
! 𝑛𝑟
tics from snapshots can be found at https://github.com/ 𝑻 −1 ©∑︁ ∗ 𝑻 −1 ∗ ª
+ 𝒙𝑖 𝚺 𝒙𝑖 + 𝒙 𝑗 𝚺 𝒙 𝑗® , (A1)
franciscovillaescusa/Quijote-simulations. Additionally, 𝑖=1 « 𝑗=1 ¬
a Python3 package with code examples and documentation is pro-
vided at https://github.com/CompiledAtBirth/pyCARPool E-step. Consider conditional expectation of the log-likelihood over
to experiment with CARPool. missing data 𝒔∗ given the (observed) data and the covariance at the
Figure 6. Confidence contours of the cosmological parameters computed using the Fisher matrix based on the estimated matter correlation function covariance
matrix. The estimators which we compare are the same as in Figure 1.
𝑘-th step, 𝚺[𝑘] We stress that equations (A4), (A5) and (A6) depend on 𝑘 because
− 2E𝒔∗ |𝒓 ∗ ln L ({𝒙}, {𝒙 ∗ }|𝚺[𝑘]) = (𝑛𝑠 + 𝑛𝑟 ) ln [det (𝚺[𝑘])]
we use 𝚺[𝑘] as 𝚺.
! Writing
𝑛𝑠
© 𝑛𝑟 ∗ 𝑻
∑︁ ∑︁
𝑻 −1 −1 ∗ ª
+ 𝒙 𝑖 𝚺[𝑘] 𝒙 𝑖 + E𝒔∗ |𝒓 ∗
𝒙 𝑗 𝚺[𝑘] 𝒙 𝑗 ® + 𝑐 𝑛𝑠
∑︁
𝑖=1 𝑗=1
« ¬
𝑛𝑠 𝚺
b= 𝒙 𝑖 𝒙𝑻𝑖
Using linearity of expectation we can look at each summand of 𝑖=1
Figure 8. Same plot as in Figure 2 for the matter correlation function, still
𝑛𝑟 𝑨𝒓𝒓 + (𝑛𝑠 + 𝑛 𝑝 )b 𝚫
𝚺𝒓𝒓
with 𝚿emp . EM
𝚺𝒓𝒓
b = (A12)
𝑛 𝑠 + 𝑛𝑟 + 𝑛 𝑝
EM 𝚫
M-step. Maximizing the expected log-likelihood, Eq. (A7) to find 𝚺
d 𝒔𝒓 = (𝑛𝑠 + 𝑛 𝑝 )b
𝚺𝒔𝒓 (A13)
the next value of the parameter is now trivial: EM −1
× (𝑛𝑠 + 𝑛𝑟 + 𝑛 𝑝 )1 𝑝𝑠 − [ 𝚺
d 𝑟𝑟 ] 𝑛𝑟 𝑨𝒓𝒓 (A14)
1 b
EM EM −1
𝚺[𝑘 + 1] = 𝑛𝑠 𝚺 + 𝑛𝑟 𝑨[𝑘] (A8) bEM = 𝚺
𝑩 [𝚺 ] (A15)
𝑛 𝑠 + 𝑛𝑟 d 𝒔𝒓 d 𝑟𝑟
𝚫 +𝑛 𝑩 EM ( 𝑨 − b EM EM𝑇
EM (𝑛𝑠 + 𝑛 𝑝 )b
𝚺𝒔𝒓 𝑟b 𝒓𝒓 𝚺𝒓𝒓 )𝑩
b
𝚺
d 𝒔𝒔 = (A16)
A1.2 Inclusion of an Inverse-Wishart prior for 𝚺 𝑛𝑠 + 𝑛 𝑝
The generalization to maximizing the posterior for 𝚺 with a conjugate Equation (A16) is equivalent to equation (17), even though it looks
prior taking the Inverse-Wishart form is immediate. Taking 𝚿 to be more complicated. This is because solving for the EM introduces the
the parameter of the prior, 𝑃 = 2 dim(𝑠), and 𝜈 > 𝑃 − 1 the number covariance of the unpaired surrogates only 𝑨𝒓𝒓 and not the covariance
of degrees of freedom, then this modifies the EM update, Eq. (A8) of the paired and unpaired surrogates b ★.
𝚺𝒓𝒓
Δ 𝜕 ln[ P (𝚺 | {𝒙 }, {𝒓 ∗ }) ]
𝚫 b𝑇 .
𝚺𝒔𝒔
b +B bMAP 𝚺𝑟𝑟 − 𝚺
b𝑟𝑟 BMAP (A18) We then solve successively 𝜕𝛼 = 0 for 𝛼 ∈
𝚺𝒓𝒓 , 𝑩𝑻 , 𝚺𝒔|𝒓 , after a bit of derivation and linear algebra, and
we find the solutions from section 3.3.2, i.e. equations (19), (20) and
MAP
(21) which allow to compute b 𝚺𝒔𝒔 from equation (22)
correlated.
(ii) 𝚺𝒓𝒓 𝚺𝒔 |𝒓 .
(iii) 𝚺𝒓𝒓 ∼ W −1 (𝚿𝒓𝒓 , 𝜈 − 𝑝 𝑠 ).
(iv) 𝚺𝒔 |𝒓 ∼ W −1 (𝚿𝒔 |𝒓 , 𝜈) with 𝚿𝒔 |𝒓 ≡ (𝚿/𝚿𝒓𝒓 ). C2 MAP on the regression parameters
(v) 𝑩𝑻 |𝚺𝒔 |𝒓 ∼ MN (𝚿𝒓𝒓 −1 𝚿𝒓 𝒔 , 𝚺𝒔 |𝒓 ⊗ 𝚿𝒓𝒓 −1 ) where ⊗ is the
Kronecker product and MN designates the matrix normal distribu- We chose to present one particular example of the MAP estimate
tion. from section 3.3.2 on the power spectrum, which showed the most
successful results with the "block" parametrization from section 3.3.1
This is particularly convenient for our problem and we can extend An- with only 𝑛𝑠 = 10 + 5 simulations. We fix 𝜈 = 2 ∗ 𝑝 𝑠 + 2 in this
derson’s result straightforwardly to a Maximum A Posteriori (MAP) case and do not consider it a free parameter, nor do we allow it
estimate. In particular, we can re-parametrize the distribution to define an improper prior, i.e. we do not allow 𝜈 ≤ 2 ∗ 𝑝𝑠 − 1.
This corresponds to the lowest integer for which the expectation of
P (𝚺) = P (𝑩𝑻 |𝚺𝒔 |𝒓 )P (𝚺𝒔 |𝒓 )P (𝚺𝒓𝒓 ) . (B1) the Inverse-Wishart exists. In Figure C2, the marginal confidence
bounds are much wider than the truth for both 𝑛𝑠 = 10 and 𝑛𝑠 =
160 for the CARPool Bayes estimator (this time the "regression"
Let’s index the unpaired surrogate samples as 𝒓 𝑖 with 𝑖 = 1, . . . , 𝑛𝑟 framework from section 3.3.2). Since the MAP on the regression
and the surrogate samples that are part of the pairs 𝒙 𝑖 𝑖 = 1, . . . , 𝑛𝑠 parameters does not allow for an improper prior, the estimator of the
as 𝒓 𝑖 with 𝑖 = 𝑛𝑟 + 1, . . . , 𝑛𝑟 + 𝑛𝑠 . We factorize the likelihood as simulation covariance puts too much weight on the naïve diagonal
Anderson, that is to say empirical Bayes prior we use (section 3.4.2). For future studies, we
can explore whether having a "smarter prior", for instance a model
𝑛𝑠 𝑛𝑟
Ö Ö covariance computed from theoretical approximations to parametrize
L ({𝒙}, {𝒓 ∗ }|𝚺) = P (𝒙 𝑖 | 𝝁 𝒔 , 𝚺𝒔𝒔 ) P (𝒓 𝑗 | 𝝁𝒓 , 𝚺𝒓𝒓 ) (B2)
the Inverse-Wishart distribution, can significantly improve or not both
𝑖=1 𝑗=1
the CARPool Bayes estimators from sections 3.3.1 and 3.3.2.
𝑠 +𝑛𝑟
𝑛Ö 𝑛𝑠
Ö
= P (𝒓 𝑖 | 𝝁𝒓 , 𝚺𝒓𝒓 ) P (𝒔𝑖 | 𝝁 𝒔𝑖 |𝒓 𝑖 , 𝚺𝒔 |𝒓 )
𝑖=1 𝑖=1
C3 Results from the bispectrum
The right hand side depends separately on 𝚺𝒓𝒓 , 𝚺𝒔 |𝒓 and 𝑩𝑻 (through Here, we directly present the confidence bounds for the ΛCDM pa-
𝝁 𝒔 |𝒓 ) as the prior, so we can solve the MAP problem from equation rameter found using various covariance estimators of the bispectrum.
(12). The motivation here is to demonstrate the improvement over CW21
Then the natural logarithm posterior distribution is for the same summary statistics. The first summary statistics we
Figure C1. Fisher confidence contours of the cosmological parameters based on the estimated covariance matrix of the matter power spectrum. The estimators
which we compare are the same as in Figure 1, except that we have now 𝑛𝑠 = 40 + 10 simulations for CARPool Bayes (empirical Bayes prior on the block
covariance).
test is the set of squeezed isosceles triangles, that is to say the bis- REFERENCES
pectra computed for 𝑘 1 = 𝑘 2 and in ascending order of the ratio
Alsing J., Wandelt B., 2018, MNRAS, 476, L60
𝑘 3 /𝑘 1 ≤ 0.20 (𝑝 𝑠 = 98 in this case).
Alsing J., et al., 2020, ApJS, 249, 5
Alves de Oliveira R., Li Y., Villaescusa-Navarro F., Ho S., Spergel D. N.,
2020, arXiv e-prints, p. arXiv:2012.00240
Figure C3 demonstrate that we get parameter constraints much Anderson T. W., 1957, Journal of the American Statistical Association, 52,
more representative of the truth with 𝑛𝑠 = 20 + 10 simulations that 200
with the sample covariance using 𝑛𝑠 = 110 simulations. The CAR- Angulo R. E., Pontzen A., 2016, MNRAS, 462, L1
pool Bayes estimator is the one from section 3.3.1 using the empirical Angulo R. E., Zennaro M., Contreras S., Aricò G., Pellejero-Ibañez M.,
Bayes prior from section 3.4.2. Stücker J., 2020, arXiv e-prints, p. arXiv:2004.06245
Bai Z. D., Yin Y. Q., 1993, The Annals of Probability, 21, 1275
Bernardeau F., Colombi S., Gaztanaga E., Scoccimarro R., 2002, Phys. Rept.,
367, 1
Then, we take a look at the reduced bispectrum of equilateral Blot L., Corasaniti P. S., Alimi J.-M., Reverdy V., Rasera Y., 2014, Monthly
equilateral triangles with 𝑘 1 = 𝑘 2 = 𝑘 3 varying up to 𝑘 max = 0.75 Notices of the Royal Astronomical Society, 446, 1756
ℎMpc−1 (𝑝 𝑠 = 40). In Figure C4, we observe the CARPool Bayes Blot L., Corasaniti P. S., Amendola L., Kitching T. D., 2016, MNRAS, 458,
estimator gives almost identical parameters marginal contours to the 4462
truth with only 𝑛𝑠 = 10 + 5 simulations, while the sample covariance Blot L., et al., 2019, MNRAS, 485, 2806
of simulations uses 𝑛𝑠 = 100 simulations. Chartier N., Wandelt B. D., 2021, Monthly Notices of the Royal Astronomical
Figure C2. Confidence contours of the cosmological parameters computed using the Fisher matrix based on the estimated matter power spectrum covariance
The "CARPool Bayes" estimates follow the computations of section 3.3.2, where the prior, still the empirical Bayes one from section 3.4.2, is parametrized
given the regression parameters. We stress that this is the only Figure in the paper that shows a computation of the "regression" MAP from section 3.3.2.
Society, 509, 2220 Friedrich O., Eifler T., 2018, MNRAS, 473, 4150
Chartier N., Wandelt B., Akrami Y., Villaescusa-Navarro F., 2021, Monthly Friedrich O., et al., 2021, MNRAS,
Notices of the Royal Astronomical Society, 503, 1897 Gallier J., 2011, Geometric Methods and Applications. Springer New
Cheng S., Yu H.-R., Inman D., Liao Q., Wu Q., Lin J., 2020, arXiv e-prints, York, doi:10.1007/978-1-4419-9961-0, https://doi.org/10.1007%
p. arXiv:2003.03931 2F978-1-4419-9961-0
Chuang C.-H., Kitaura F.-S., Prada F., Zhao C., Yepes G., 2015, MNRAS, Garrison L., 2019, PhD thesis, University Of Washington
446, 2621 Garrison L. H., Eisenstein D. J., Ferrer D., Tinker J. L., Pinto P. A., Weinberg
Colavincenzo M., et al., 2019, MNRAS, 482, 4883 D. H., 2018, ApJS, 236, 43
Dai B., Seljak U., 2020, arXiv e-prints, p. arXiv:2010.02926 Giocoli C., et al., 2021, A&A, 653, A19
DeRose J., et al., 2019, The Astrophysical Journal, 875, 69 Habib S., et al., 2016, New Astron., 42, 49
Dempster A. P., Laird N. M., Rubin D. B., 1977, Journal of the Royal Statistical Hall A., Taylor A., 2019, MNRAS, 483, 189
Society. Series B (Methodological), 39, 1 Harnois-Déraps J., Pen U.-L., 2013, MNRAS, 431, 3349
Desjacques V., Jeong D., Schmidt F., 2018, Phys. Rept., 733, 1 Harnois-Déraps J., Vafaei S., Van Waerbeke L., 2012, MNRAS, 426, 1262
Ding Z., et al., 2022, arXiv e-prints, p. arXiv:2202.06074 Harnois-Déraps J., Pen U.-L., Iliev I. T., Merz H., Emberson J. D., Desjacques
Dodelson S., Schneider M. D., 2013, Phys. Rev. D, 88, 063537 V., 2013, MNRAS, 436, 540
Eifler, T. Schneider, P. Hartlap, J. 2009, A&A, 502, 721 Harnois-Déraps J., Giblin B., Joachimi B., 2019, A&A, 631, A160
Escoffier S., et al., 2016, arXiv e-prints, p. arXiv:1606.00233 Hartlap J., Simon P., Schneider P., 2007, A&A, 464, 399
Favole G., Granett B. R., Silva Lafaurie J., Sapone D., 2020, arXiv e-prints, Hassan S., et al., 2021, arXiv e-prints, p. arXiv:2110.02983
p. arXiv:2004.13436 He S., Li Y., Feng Y., Ho S., Ravanbakhsh S., Chen W., Póczos B., 2019,
Feng Y., Chu M.-Y., Seljak U., McDonald P., 2016, MNRAS, 463, 2273 Proceedings of the National Academy of Sciences, 116, 13825–13832
Figure C3. Confidence contours of the cosmological parameters computed using the Fisher matrix based on the estimated matter bispectrum covariance matrix,
for a set of squeezed isosceles triangles. The estimators result from the same computations as in Figure 1
.
Heavens A. F., Jimenez R., Lahav O., 2000, MNRAS, 317, 965 Lippich M., et al., 2019, MNRAS, 482, 1786
Hikage C., Takahashi R., Koyama K., 2020, Phys. Rev. D, 102, 083514 Lucie-Smith L., Peiris H. V., Pontzen A., 2019, MNRAS, 490, 331
Howlett C., Manera M., Percival W., 2015, Astronomy and Computing, 12, Lucie-Smith L., Peiris H. V., Pontzen A., Nord B., Thiyagalingam J., 2020,
109–126 arXiv e-prints, p. arXiv:2011.10577
Ishiyama T., Fukushige T., Makino J., 2009, PASJ, 61, 1319 Maksimova N. A., Garrison L. H., Eisenstein D. J., Hadzhiyska B., Bose S.,
Izard A., Crocce M., Fosalba P., 2016, Monthly Notices of the Royal Astro- Satterthwaite T. P., 2021, MNRAS, 508, 4017
nomical Society, 459, 2327–2341 McClintock T., et al., 2019a, arXiv e-prints, p. arXiv:1907.13167
Joachimi B., 2017, MNRAS, 466, L83 McClintock T., et al., 2019b, The Astrophysical Journal, 872, 53
Joachimi B., Taylor A., 2014, Statistical Challenges in 21st Century Cosmol- Modi C., Lanusse F., Seljak U., 2020, arXiv e-prints, p. arXiv:2010.11847
ogy, 306, 99 Modi C., Lanusse F., Seljak U., Spergel D. N., Perreault-Levasseur L., 2021,
Kasim M. F., et al., 2020, arXiv e-prints, p. arXiv:2001.08055 arXiv e-prints, p. arXiv:2104.12864
Kaufman G. M., 1967, Center for Operations Research and Econometrics Mohammed I., Seljak U., 2014, Mon. Not. Roy. Astron. Soc., 445, 3382
Report no. 6710. Catholic University of Louvain. Heverlee, Belgium. Mohammed I., Seljak U., Vlah Z., 2017, MNRAS, 466, 780
Kitaura F. S., Yepes G., Prada F., 2014, MNRAS, 439, L21 Monaco P., Sefusatti E., Borgani S., Crocce M., Fosalba P., Sheth R. K.,
Kodi Ramanah D., Charnock T., Villaescusa-Navarro F., Wandelt B. D., 2020, Theuns T., 2013, MNRAS, 433, 2389
Monthly Notices of the Royal Astronomical Society, 495, 4227–4236 Muirhead R. J., 1982, Aspects of Multivariate Statistical Theory. John Wiley
Leclercq F., Faure B., Lavaux G., Wandelt B. D., Jaffe A. H., Heavens A. F., & Sons, Inc
Percival W. J., 2020, A&A, 639, A91 Paz D. J., Sánchez A. G., 2015, Monthly Notices of the Royal Astronomical
Li Y., Singh S., Yu B., Feng Y., Seljak U., 2019, J. Cosmology Astropart. Society, 454, 4326
Phys., 2019, 016 Pearson D. W., Samushia L., 2016, MNRAS, 457, 993
Figure C4. Confidence contours of the cosmological parameters computed using the Fisher matrix based on the estimated matter bispectrum covariance matrix,
for a set of equilateral triangles. The estimators we compare are the same as in Figure 1
Pedersen C., Font-Ribera A., Rogers K. K., McDonald P., Peiris H. V., Pontzen Smith A., de Mattia A., Burtin E., Chuang C.-H., Zhao C., 2021, MNRAS,
A., Slosar A., 2021, J. Cosmology Astropart. Phys., 2021, 033 500, 259
Percival W. J., et al., 2014, MNRAS, 439, 2531 Springel V., 2005, MNRAS, 364, 1105
Percival W. J., Friedrich O., Sellentin E., Heavens A., 2021, arXiv e-prints, Spurio Mancini A., Piras D., Alsing J., Joachimi B., Hobson M. P., 2022,
p. arXiv:2108.10402 MNRAS, 511, 1771
Philcox O. H. E., Eisenstein D. J., 2019, MNRAS, 490, 5931 Taffoni G., Monaco P., Theuns T., 2002, Monthly Notices of the Royal Astro-
Philcox O. H. E., Eisenstein D. J., O’Connell R., Wiegand A., 2020, MNRAS, nomical Society, 333, 623
491, 3290 Takahashi R., et al., 2009, ApJ, 700, 479
Philcox O. H. E., Ivanov M. M., Zaldarriaga M., Simonović M., Schmittfull Tassev S., Zaldarriaga M., 2012, J. Cosmology Astropart. Phys., 2012, 013
M., 2021, Phys. Rev. D, 103, 043508 Tassev S., Zaldarriaga M., Eisenstein D. J., 2013, Journal of Cosmology and
Pontzen A., Slosar A., Roth N., Peiris H. V., 2016, Phys. Rev. D, 93, 103519 Astroparticle Physics, 2013, 036–036
Pope A. C., Szapudi I., 2008, MNRAS, 389, 766 Tassev S., Eisenstein D. J., Wand elt B. D., Zaldarriaga M., 2015, arXiv
Potter D., Stadel J., Teyssier R., 2017, Computational Astrophysics and Cos- e-prints, p. arXiv:1502.07751
mology, 4, 2 Taylor A., Joachimi B., 2014, MNRAS, 442, 2728
Remy B., Lanusse F., Ramzi Z., Liu J., Jeffrey N., Starck J.-L., 2020, arXiv Taylor A., Joachimi B., Kitching T., 2013, MNRAS, 432, 1928
e-prints, p. arXiv:2011.08271 Villaescusa-Navarro F., et al., 2018, The Astrophysical Journal, 867, 137
Rogers K. K., Peiris H. V., 2021, Phys. Rev. D, 103, 043526 Villaescusa-Navarro F., et al., 2020, ApJS, 250, 2
Schäfer J., Strimmer K., 2005, Statistical applications in genetics and molec- Villaescusa-Navarro F., et al., 2021a, arXiv e-prints, p. arXiv:2109.09747
ular biology, 4, Article32 Villaescusa-Navarro F., et al., 2021b, arXiv e-prints, p. arXiv:2109.10915
Scoccimarro R., Sheth R. K., 2002, MNRAS, 329, 629 Villaescusa-Navarro F., et al., 2021c, ApJ, 915, 71
Sellentin E., Heavens A. F., 2018, MNRAS, 473, 2355 Wadekar D., Ivanov M. M., Scoccimarro R., 2020, Phys. Rev. D, 102, 123521
This paper has been typeset from a TEX/LATEX file prepared by the author.